RB Exome Manuscript

0.2.1 Samples

Untreated retinoblastoma tumor samples were obtained with informed consent under approved IRB protocols immediately following enucleation. One fraction of tumor was rapidly frozen in liquid nitrogen. A second fraction was placed in culture to establish cell lines. Matched normal DNA samples were obtained from mesenchymal cells isolated from orbital fat pad tissue collected during enucleation.

0.2.2 Establishing and propagating retinoblastoma cell lines

Dissociated retinoblastoma cells were prepared after culturing freshly explanted tumors in RB culture medium consisting of IMDM (Mediatech Inc.) with 10% FBS (Sigma-Aldrich), 2mM glutamine (Mediatech Inc), 55uΜ beta-mercaptoethanol (Sigma-Aldrich), and 10ug/ml insulin (Humulin-R, Eli Lilly and Company) at 37o C in a humidified incubator with 5% CO2 [Xu, 2013].

0.2.3 DNA isolation and Sequencing

Flash-frozen tumor samples were dissociated by triteration and immediately prepared for DNA extraction. In the same fashion, cryopreserved cell lines and normal MSC samples were dissociated and DNA was isolated using the Qiaamp DNA Minikit (Qiagen, Hilden, Germany). A minimum of 1 ug of DNA was submitted for DNA Exome Sequencing on Illumina HiSeq by Macrogen Inc. (Seoul, South Korea) and BGI (Shenzhen, China). A sequencing depth of 150x for cell line and matched normal samples and 300x for tumor samples was requested.

0.2.4 Methods-Variant Calling Pipeline

Exome capture sites were defined according to the Sureselect V5 Capture Library (Agilent). Sequences were aligned to USSC hg19 human reference genome using bwa (Li and Durbin 2010). Following the GATK best practices workflow for somatic variant discovery, duplicate reads were removed and quality scores were recalibrated. Variant calling was done using GATK Mutect2 for somatic variants between tumor and matched-normal samples.(DePristo et al. 2011; Van der Auwera et al. 2013). Putative germline variants were assessed by assembling a panel of normals with mutect2 in variant detection mode.

0.2.5 Methods-Variant Filtration

Variants were removed if they were documented in the Genome Aggregation Database (gnomAD) Lek et al. (2016) at an allele frequency greater than ten percent . All admitted variants were marked “PASS” by GATK. Variants were removed if they appeared in the matched normal at greater than Low frequency (<5% VAF ???) . Variants were also filtered out if they appeared in a panel of normals at Low frequency (<5% VAF ???) . Two variants were considered identical when they had identical genomic coordinates, reference and mutant allele sequences, and GATK genotype calls (AA, AB, BB).

0.2.6 Somatic Copy Number Alteration

Somatic copy number alteration (SCNA) were detected using the R package CopywriteR which uses off-target reads in targeted sequencing experiments (Kuilman et al. 2015).

0.2.7 Allelic Imbalance

Variant sites detected by Mutect2 in any sample were collected into a single .bed file. B allele frequencies for all samples were calculated at these sites. B allele frequency segmentation was performed as described Staaf et al. (2008), yielding a mirrored B allele frequencey(mBAF) representing an abnormal B allele frequency over a given region. A region was designated as having allelic imbalance if mBAF was greater than 0.56 and was designated as having loss of heterozygosity (LOH) if mBAF was greater than 0.7.

0.3.1 Cell Line Growth Curves

Figure 1: Retinoblastoma cell line growth curves

Figure 1: Retinoblastoma cell line growth curves

Retinoblastoma cell lines were established from enucleated tumor prior to treatment, then expanded in culture under either of two protocols (Reynolds and VC) and viably frozen at a minimum cell number. For this study, these initial cell lines were thawed and passaged on a schedule of 1-2 days upon color change of phenol red to amber. With an intial volume of 200 ul cell solution per 96 well plate, cells were split 1:2 wells following successive days of sufficient color change. A log of well number based on volume of a 96-well plate was maintained for each cell line, with the first value corresponding to the cumulative days in culture up to initial date of thawing, and correcting for cells banked in liquid nitrogen storage at regular intervals. Indicate range of cell # used to start cultures; All are pre-treatment , CHLA, established by either of two protocols (Reynolds and KS)

0.3.2 RB patient characteristics

[1] “Table 1: Patient Characteristics” Clinical characteristics of study participants. Clinical sequencing records for four of twelve samples are available in supplementary info, all rb1 clinical variants were recapitulated by exome sequencing with the exception of biallelic promoter methylation reported in sample T-20.

0.3.3 RB1 Status in Tumor and Cell Line Samples

[1] “Table 2: RB1 and MCYNA status in Tumor and Cell Line Samples” Samples are displayed by tumor and cell line id. The estimated number of RB1 alleles subject to SCNA loss is displayed for each sample along with detected LOF mutations to RB with variant allele frequency (VAF). Also displayed are determinations of copy-neutral loss of heterozygosity (CN-LOH) with mBAF spanning RB. If a copy number gain spanning MYCNA was detected, the sample is labeled MYCNA. In most cases RB1 initiation could be accounted for by the sum of alterations in RB1 or through MYCNA, with the exception of samples 20-T and 20-CL. These samples were determined to be subject to RB1 promoter methylation by clinical sequencing.

0.3.4 Somatic Variants

Following variant calling, a total of 643,581 variants were discovered. After filtering out variants that were found in the panel of matched normal controls, 16,348 (16,348/643,581; 2.5%) variants were recorded. After further filtering by pathogenicity, functional annotation, and variant allele frequency (VAF) in gnomAD, a final tally of 220 variants were found (679/16,348; 4%). Details can be found in Table 3.

0.3.5 Somatic Variants Shared Between Tumor and Cell Line

[1] "**Table 3: Variants Recurrrent in Tumor and Cell Line Samples" Somatic variants shared between matched retinoblastoma tumor and cell line samples included RB1, PAN2, EP300, and , recurrent in six of 12 samples. Additional SNVs including BCOR, previously identified Kooi, Mol, Massink, Ameziane, et al. (2016), Zhang et al. (2012) and NAF1 were not detected in cell line samples but were found in tumor. Recurrent variants included RB1 and EP300, closely related to CREBBP identified by Kooi, Mol, Massink, Ameziane, et al. (2016). Additional somatic variants exclusive to tumor and cell line samples can be found in supplementary info S1.

Indicate TTL5 changes as one variant Table with allele (Discussion: BCOR mutation in T28 ‘lost’ in CL 28; suggesting does not help and might hurt growth. 5 of 7 other CNAs were retained (all except 12-) suggesting that the rare BCOR wild type and chr12 normal cell was selected; suggesting that 12- and BCOR mutation were the last changes in the evolution of the tumor (consisstent with BCOR VAF = 0.394); suggesting either selection for two copies of chr 12 or for wild type BCOR. [also check CL28 for CN-LOH] (similar points for NAF1)

PAN2 AKAP12: mention increase VAF in cell lines – though not much can be made of it. DO indicate likely effects of mutation on protein structure and function. If they are heterozygous in cancer cell may be dominant acting oncogenic or haplo-insufficent if mutation inactivates a tumor suppressor function. Discussion: note prior evidence for role in Cancer (AKAP12) but no prior papers for PAN2; how affects RNA metabolism/processing it may be consistent with Kooi observation that progression is associated with increased RNA biogenesis??? Are EP300 ‘mutations’ also found at very low frequency in other tumors, cell lines, normals in our cohort.

0.3.6 Somatic Copy Number Alterations

Somatic copy number alterations (SCNA) were detected in all twelve samples. Regions of copy number change included 1q, 2p, 6p, 7q and 13q–conistent with (Kooi, Mol, Massink, Jong, et al. 2016). We detected amplification of 2p in the region spanning the MYCN oncogene in sample 24-T and matched cell line. We also found SCNA loss in sample 14-T on the X chromosome covering BCOR, previously implicated in retinoblastoma pathogenesis(Zhang et al. 2012, @kooi_somatic_2016).

Somatic Copy Number Alterations

Somatic Copy Number Alterations

0.3.7 SCNA in Kooi peak regions

[1] “SCNA overlap with Kooi et. al peak regions in vision center tumors and cell lines”

[1] “SCNA overlap with Kooi et. al peak regions in reynolds cell lines”

0.3.8 Allelic Imbalance Loss of Heterozygosity

Regions of loss of heterozygosity with a threshold of 0.7 mBAF are shown in supplementary info.

Regions of LOH

Regions of LOH

Regions of LOH

0.3.9 Germline Variants

No germline RB1 mutations were detected for the 10 tumors from unilaterally affected patients, consistent with these patients having acquired two somatic RB1 alterations.
One bilateral pt had 13q focal deletion that included the RB1 gene. In addition, in a second bilateral case (VCRBVC41) we did not detect a germline variant. However, the tumor showed RB1 LOH, there is likely to be a hidden germline RB1 abnormality, such as promoter mutation (ref) or deep intron mutation (ref). As a less likely possibility, this patient could have a low-level mosaic mutation or promoter methylation that was not detected in the MSC DNA. However, mosaicism seems unlikely to yield bilateral tumors. (let’s check for low VAF changes in normal DNA). [check Gallie/Lohmann paper and Rushlow MYCNA paper might mention other alternatives to explain lack of RB1 mutation detection)

0.4.1 RB1 Hits

The number of

1. Previous studies have found that somatic genomic alterations are rare in retinoblastoma (Kooi, Mol, Massink, Ameziane, et al. 2016). Our findings are that retinoblastoma variants are almost exclusively within the RB1 gene. We did not find a dramatic increase in recurrent copy number changes despite our aim for increased sequencing depth. The most common variants

0.4.2 Supplementary Info Somatic Variants in Vision Center Tumors

[1] “variants in tumors”

0.4.3 Supplementary Info Somatic Variants in Vision Center Cell Lines

[1] “variants in cell lines”

0.4.4 Supplementary Info Somatic Variants in Reynolds Cell Lines

[1] “variants in cell lines”